Improved stereochemical analysis of conformationally flexible diamines by binding to a bisporphyrin molecular clip

Article abstract of DOI:10.1002/chem.201300533

The relative stereochemistry of acyclic diamines with several stereogenic centers has been analyzed by NMR spectroscopy in combination with conformational deconvolution. Binding to a bisporphyrin molecular clip improves the stereochemical assignment signi

Full text of DOI:10.1002/chem.201300533

FULL PAPER
DOI: 10.1002/chem.201300533
Improved Stereochemical Analysis of Conformationally Flexible Diamines by
Binding to a Bisporphyrin Molecular Clip
Sara Norrehed, Henrik Johansson, Helena Grennberg, and Adolf Gogoll*[a]
Abstract: The relative stereochemistry of acyclic diamines with several stereogenic
centers has been analyzed by NMR spectroscopy in combination with conforma-
tional deconvolution. Binding to a bisporphyrin molecular clip improves the ster-
eochemical assignment significantly. The diamines were synthesized from inexpen-
sive sugar alcohols, and their stable hydrochlorides were quantitatively converted
into free bases by treatment with ion-exchange resin.
Keywords: configuration determi-
nation conformation analysis
host–guest systems · N ligands ·
·
·
NMR spectroscopy
Introduction
the nuclear Overhauser effect (NOE)). The method has
been successfully applied to small and medium-sized mole-
cules.[4]
Whenever chemical syntheses result in the formation of
molecules with several stereogenic centers, it is crucial to
determine their relative stereochemistry. In contrast to rigid
molecules, or those with only a few conformers, this can be
quite difficult to achieve for small, flexible molecules. In this
situation, NMR spectroscopic parameters are obtained as
population-weighted time averages and cannot truly be rep-
resented by a single conformation. To facilitate assignments
under these preconditions, several methods are in common
use.[1] Whereas coupling constant (J)-based configurational
analysis assumes the data is dominated by one or a few con-
formers, Kishiꢀs universal NMR database (UDB) method
matches experimental spectral parameters against a library
of model compounds without involving conformational anal-
ysis.[2] Computational methods for the prediction of con-
former distributions and hence the resulting NMR spectro-
scopic parameters have to rely on a detailed knowledge of
molecular interactions in solution, which might be difficult
to achieve. They are therefore very successful for molecules
for which a single conformation is dominant, such as pro-
teins, but not as much for small, flexible molecules.
While all these approaches handle a scenario in which the
organic molecule is present free in solution, we found it
tempting to investigate whether conformational restriction
imposed on the flexible molecule might improve such an
analysis. Diamines bound to metalloporphyrins emerged as
suitable systems for this investigation, since their binding in-
teraction has been studied extensively.[5–7] In particular,
binding to bisporphyrin tweezers is responsive to a variety
of molecular properties. Examples include differentiation
between guests of varying size,[6] and in particular determi-
nation of the absolute stereochemistry of guest molecules,
often through the circular dichroism of the formed com-
plexes.[7] Usually, the conformations of guest molecules in
these complexes are analyzed in terms of a single, dominat-
ing conformation, or are undetermined. For our purpose, we
required a rigid, symmetric bisporphyrin molecular clip.
Such a host would have to form host–guest complexes with
diamines, have a small number of NMR spectroscopic sig-
nals, and also provide chemical-shift effects due to the dia-
magnetic porphyrin system, thus minimizing signal overlap
between host and guest. We therefore chose the glycoluril
bisporphyrin clip 1 (Schemes 1 and 2).[8]
A different approach is used in NMR spectroscopic analy-
sis of molecular flexibility in solution (NAMFIS), in which a
complete set of conformers by matching against experimen-
tal NMR spectroscopic data is reduced to a subset, eventual-
ly obtaining a conformer population that is likely to gener-
ate the observed parameters.[3] Here, one can chose a pa-
rameter that, in contrast to scalar coupling constants (J) and
chemical shifts (d), is not biased by substituent effects (i.e.,
Suitable guest molecules with known stereochemistry
were prepared from two alditols by conversion into a,w-dia-
minopolyolmethoxy ethers. Sorbitol was transformed into
[a] S. Norrehed, Dr. H. Johansson, Prof. Dr. H. Grennberg,
Prof. Dr. A. Gogoll
Uppsala University, Department of Chemistry
BMC, Box 576, 751 23 Uppsala (Sweden)
E-mail: adolf.gogoll@kemi.uu.se
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201300533.
Scheme 1. Bisporphyrin clip 1 with a glycoluril scaffold.
Chem. Eur. J. 2013, 19, 14631 – 14638
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
14631

Scheme 4. Synthesis of 1,w-diaminodideoxyalditol methyl ethers (here
starting from sorbitol). a) TrCl, pyridine, 1008C, 69%; b) MeI, DMF,
NaOH, 77%; c) p-TSA monohydrate, MeOH/CH2Cl2, 77%; d) Et3N,
MsCl, 99%; e) NaN3, DMF, 708C, 70%; f) chromatographic materials,
70%; g) LiAlH4 (1), HCl (2), 08C, 89%; h) Amberlite IRA-400, MeOH,
quantitative.
Scheme 2. The binding of (2S,3R,4R,5R)-1,6-diamino-2,3,4,5-tetramethox-
yhexane (2) and clip 1 is illustrated.
compound 2, which has no symmetry elements, and xylitol
provided meso compound 3. In these guests, methoxylation
of the hydroxy groups was expected to prevent the confor-
mational equilibria to be dominated by hydrogen-bonding
interactions.
Sorbitol 2a was tritylated at the terminal positions with
tritylchloride/pyridine (69%), and the resulting tetraol 2b
was methylated with iodomethane to afford the tetramethyl
ether 2c (77%). Detritylation (p-TsOH in MeOH/CH2Cl2)
yielded the tetramethylated sorbitol 2d (77%). Mesylation
with mesyl chloride afforded 2e (crude yield from NMR
spectroscopy integration 99%). Reaction of 2e with sodium
azide gave the diazide 2g (yield 70%). Reduction of 2g
with LiAlH4 produced the diamine, which after treatment
with dry hydrogen chloride was obtained as the hydrochlor-
ide 2h (yield 89% from diazide). Total yield from 2e: 62%.
In an analogue synthesis, xylitol was converted into the
(2R,3R,4S)-1,5-diamino-1,5-dideoxy-2,3,4-tri-O-methylxylitol
dihydrochloride (3), 39% yield from mesylate 3e (see the
Supporting Information for details).
Results and Discussion
Synthesis of diaminopolyol methoxy ethers: Diaminodideox-
yalditols constitute an interesting group of monomers for
the preparation of biodegradable polymers from renewable
starting materials.[9,10] In a previously published synthetic
route starting from a 1-benzyl hexopyranoside,[11] a sequence
of methylation with MeI, debenzylation, reduction, mesyla-
tion, nucleophilic substitution, and subsequent reduction of
the diazide to diamine afforded the title compounds in
seven steps. One peculiar observation was the formation of
cyclic ethers from the mesylated intermediates.[10,11] Initially,
a series of alternative, apparently simpler synthetic routes
was considered. Thus, oxidation of the alditol to aldaric
acids, followed by methylation, ammonolysis to the dia-
mides, then followed by reduction to diamines yielded a
complex product mixture.[12] Sulfonylation of an alditol[13]
followed by azidation gave a mixture of products in low
yield. Therefore, we devised a strategy related to the one by
Galbis et al.[11] but starting from alditols (Schemes 3 and 4).
Deprotonation: Hydrochloride salts are the derivative of
choice for storage of amines without decomposition. How-
ever, quantitative deprotonation of the dihydrochlorides 2h
and 3h by the addition of various bases was difficult to ach-
ieve with satisfactory purity due to the solubility of the pro-
tonated species in both water and organic solvents. On the
other hand, Amberlite IRA-400 resin was found to provide
a fast, convenient, and clean method to convert 2h and 3h
into the free amines 2 and 3, by using MeOH as solvent to
prevent formation of emulsions, in quantitative yields. This
deprotonation also works equally well for very basic chelat-
ing diamines such as bispidine derivatives and for terminal
diamines with long (up to 20 carbon atoms) alkyl chains.[14]
This method has previously been used for the deprotonation
of base-sensitive oligosilsesquioxane amine hydrochlor-
ides.[15]
Formation of cyclic byproducts: During workup, the mesyl
methyl ethers 2e and 3e but not 4d were found to undergo
Scheme 3. Alditol starting materials.
14632
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 14631 – 14638

Conformationally Flexible Diamines
FULL PAPER
lated after column chromatography of 2e. To check the
impact of stereochemistry, the ribitol analogue 4e was syn-
thesized. When tested, 4e did not form any cyclic byproduct
(i.e., 4e remained intact after column chromatography and
after standing in solution over silica for 12 h). As for the
impact of the stereochemistry, it was confirmed that cycliza-
tion only occurs when the ring alkoxy groups are in a trans
position, such as in 2 f and 3 f/3 f’ as opposed to 4 f.
Scheme 5. Proposed mechanism for demethylative cycloetherification of
3e.
Binding studies of guests 2 and 3 with clip 1: Binding be-
tween clip 1 and diamines 2 and 3, respectively, was indicat-
ed by UV/Vis and by 1H NMR spectroscopy. Thus, a redshift
of the Soret band by 7 nm was obtained at 1:1 ratios.[7e]
NMR spectroscopic titration of the diamines with 1 resulted
in the expected chemical-shift change of guest signals to
lower values, with a maximum change for 1:1 molecular
ratios. Ditopic binding was also indicated by the number of
observed NMR spectroscopic signals, which were in accord-
ance with the diamines being positioned on the inside of the
clip and binding to both porphyrin walls. Furthermore, pro-
tons closest to the amino groups showed the largest chemi-
cal-shift change upon binding. At lower temperatures, some
signal broadening was observed. This indicates residual dy-
namics at higher temperatures, which eventually ceases
upon cooling (Figures S38–41 in the Supporting Informa-
tion).
Scheme 6. Possible products from ring closure of dimesyl-o-methyl deriv-
atives 2e, 3e, and 4e. Only compounds 2 f and 3 f/3 f’ were formed.
a slow demethylation to form cyclic ethers (Schemes 5 and
6). Such a demethylation has previously only been reported
by Galbis et al. (also for 3e).[11] The analogous debenzylative
cyclization is a fairly common reaction of benzylated polyols
with good leaving groups,[16] and its dependence on stereo-
chemistry has been discussed by Defaye and Horton.[17] The
mechanism involves anchimeric
1
The H NMR spectra of 2 and 3 are first order, with sig-
nals within a narrow chemical-shift range. Upon binding to
clip 1, signals become thoroughly separated (Figure 1) as is
commonly observed for organic molecules that bind to por-
phyrin hosts.
assistance of a benzyloxy group
in the d position to a good leav-
ing group by means of the for-
mation of
a cyclic oxonium
ion.[16a] Since this side reaction
reduces yield and purity of the
mesyl methyl ethers, we found
it worthwhile to look into the
conditions for this reaction to
occur.
Formation of cyclic com-
pounds occurred during the
preparative workup of 3e from
a synthesis performed in a non-
nucleophilic solvent such as
CH2Cl2. The reaction was facili-
tated by the presence of chro-
matographic materials (SiO2,
basic Al2O3, neutral Al2O3, and
Florisil). Contrary to previous
observations,[11] we found 3e to
be stable in room-temperature
solutions in the absence of such
materials. In the same manner
Figure 1. 1H NMR spectra of diamines 2 and 3 bound to clip 1. Inset (plotted at same scale): free diamines
the cyclic byproduct 2 f was iso- (500 MHz, CDCl3 solution, 258C).
Chem. Eur. J. 2013, 19, 14631 – 14638
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14633

A. Gogoll et al.
native assignments in Scheme 9
the one labeled 2-A is the cor-
rect one, it should be empha-
sized that a J-based analysis
would not be sufficient to dis-
tinguish between 2 and all of its
diastereomers, or between
and its diastereomers.
3
Signal assignment using con-
AHCTUNGTERGfNNUN ormational deconvolution of
NMR spectroscopic parame-
ters: The four alternatively as-
signed structures 2-A through
2-D were subjected to
a
NAMFIS-type analysis[3] (i.e., a
set of NMR spectroscopic pa-
rameters was fitted against
varying populations of possible
conformations). This conforma-
tional deconvolution was car-
ried out by using DISCON soft-
ware, which expresses the quali-
À
Scheme 7. Newman projections along C C single bonds in the proposed main rotamer of 2 (box) and its dia-
stereomers, identified by their parent alditol. The proton pair generating the largest J value when comparing
ty of fit in terms of root–mean–
square (RMS) values.[18] These
JACHTUNGTRENNUNG(H-2,H-3) to JACHTUNGTRENNUNG(H-4,H-5), JACHTUGTNRENNUG(H-1a,H-2) to JAHTCUNGRTEN(NUNG H-1b,H-2), or JACHTUNGTRENNG(UN H-6a,H-5) to JACHTUGNTREN(NUGN H-6b,H-5) is marked in red.
RMS values are calculated by
comparing the experimental pa-
Signal assignment: Initial signal assignment was achieved by
using J-based arguments. Whereas the sequential connectivi-
ty of H-1 through H-6 is unequivocal, it is still necessary to
distinguish between the signals of H-1a/b versus H-6a/b. If
we assume an all-trans carbon chain to represent the main
conformer, we can identify substituent arrangements about
rameter values (NOEs and J) with those calculated for the
chosen population of conformations. For the free diamine 2,
À
each C C single bond with the largest substituents in anti
positions (Scheme 7, sorbitol stereochemistry). The diaster-
eomers of sorbitol are included in the analysis to illustrate
that these J-based arguments would not clearly single out
one stereoisomer if the stereochemistry of the investigated
compound were unknown. Here, when looking at sorbitol
for the assignment of 2, the dihedral angle between the two
protons H-2 and H-3 is 608, and for H-4 and H-5 this angle
is 1808. Therefore, J
larger than J(H-2,H-3). Of the two alternative coupling con-
stants, we therefore assigned (H-4,H-5)=5.2 Hz and
(H-2,H-3)=2.7 Hz, respectively. This then results in the as-
ACHTUNGTRNE(NUNG H-4,H-5) would be expected to be
ACHTUNGTRENNUNG
JACHTUNGTRENNUNG
JACHTUNGTRENNUNG
signment of H-1 through H-6. Clearly, the difference in the
coupling constants is small, and we cannot be certain that
the rotamer shown in Scheme 7 is the dominant one.
Finally, amongst the diastereotopic methylene protons on
C-1 and C6, H-6a/H-6b are distinguished by realizing that
H-6a should have a smaller coupling with H-5 (608 dihedral
angle) than H-6b (1808 dihedral angle). Likewise, H-1a (608
dihedral angle) should have a smaller coupling than H-1b
(1808 dihedral angle) with H-2. The same argument is used
for the other diamino derivative 3 (Scheme 8) as well as in-
termediates 2b–h, 3b–h, and 4b–e. Although this analysis
makes it likely (although not certain) that of the four alter-
À
Scheme 8. Newman projections along C C single bonds in the proposed
main rotamer of 3, 4, and diastereomer, identified by their parent alditol.
The proton pair generating the largest
J
value when comparing
J
J
N
E
(H-1a,H-2) to J
G
ACHTUNGTRENNUNG
14634
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 14631 – 14638

Conformationally Flexible Diamines
FULL PAPER
Scheme 9. Alternative directional assignments of the C1–C6 chain, and of
the diastereotopic protons Ha/Hb of diamine 2 derived from sorbitol.
the analysis resulted in a small variation of RMS values (de-
rived from NOEs and J values) with 2-A providing the best
fit (Table 1). The same result was obtained for the bound di-
amine. It should be noted that the dominant conformation
of the bound 2 was not all-trans, but gauche interactions
Table 1. Assignment analysis of diamine 2 by conformational deconvolu-
tion.
NOE distance RMS
Figure 2. Main conformations of 2 bound to clip 1. Top: Dominant con-
formation (ꢀ75%). Bottom: All-trans conformation (ꢀ25%).
Assignment
(Scheme 9)
free diamine[a]
bound diamine
U
all-trans conformer[b]
population[c]
2-A
2-B
2-C
2-D
0.31
0.42
0.47
0.44
0.43
0.71
0.59
0.66
0.32
0.45
0.46
0.46
stressed, however, that a comparison of RMS values is
meaningful only within the same series of isomeric struc-
tures, since different series might have this parameter de-
rived from different parameter numbers and experimental
data.
[a] Population from unrestricted conformational search reduced by re-
dundant conformation elimination. NOEs and J values were used in the
analysis. [b] Only NOEs were used in the analysis. [c] Population of con-
formations with the N N distance restricted to 4–9 ꢂ. Only NOEs were
used in the analysis.
À
Conformational analysis: The free diamine 2, due to its high
flexibility, is expected to adopt a multitude of conformations
in solution. Thus, conformational deconvolution yielded four
main families of conformations (Figure 3) with a molar ratio
of approximately 1:1:1:2 (RMS 0.31).
In contrast, the free symmetric diamine 3 derived from
xylitol was not a suitable candidate for this analysis due to
the chemical equivalence of several protons (i.e., H-2/H-4,
H-1a/H-5a, H-1b/H-5b, and MeO-2/MeO-4). By signal aver-
aging on the NMR spectroscopic timescale, these protons,
although being nonequivalent in the various conformations,
become degenerate in chemical shift. Therefore the experi-
mental data is indeterminate for the analysis and no accept-
able matching of conformers can be obtained.
were present (Figure 2 and Scheme 10). Most likely, these
are a result of suboptimal fitting between the diamine and
the cavity provided by the clip 1. As expected, trying to fit
the data to a single, all-trans conformation, which in the
analysis corresponds to approximately 25% of the popula-
tion, resulted in increased RMS values and a modified order
of best fit assignment. Thus, conformational deconvolution
provided the correct signal assignment without requiring
presumptions about conformational preferences. It must be
In complex with 1, diamine 2 exists in a 1:3 molar ratio of
two dominant conformations (RMS=0.32). Of these, the
À
all-trans conformer (25%) had an N N distance of 8.9 ꢂ. In
contrast, the most dominant conformation (75%) exhibits a
carbon chain with a gauche conformation about one of the
À
À
C C bonds (Figure 2 and Scheme 10) and an N N distance
of only 7.6 ꢂ.
À
Scheme 10. Newman projections along C C single bonds in the dominant
conformation of bound 2.
Chem. Eur. J. 2013, 19, 14631 – 14638
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14635

A. Gogoll et al.
Table 2. Conformational deconvolution analysis of free and bound 2 and
its diastereomers.
Isomer derived from[a]
Stereochemistry[b]
RMS, free
diamine[c]
RMS, bound
diamine[d]
sorbitol
allitol
galactitol
mannitol
talitol
2S,3R,4R,5R
2R,3R,4S,5S
2R,3S,4R,5S
2R,3R,4R,5R
2R,3R,4S,5R
2R,3S,4S,5R
0.31
0.38
0.38
0.39
0.40
0.66
0.37
0.49
0.50
0.69
0.57
0.49
iditol
[a] Stereochemistry indicated by the corresponding alditol. [b] Assign-
ment in the order used for 2. [c] Population from unrestricted conforma-
tional search. [d] Single all-trans conformers.
the diamine derived from sorbitol, which is identical to the
actual isomer 2. However, the RMS variation is very small
(DRMS=0.07–0.09, with the exception of the iditol conge-
ner with DRMS=0.35). Therefore, it is debatable if this
would be considered sufficient to confirm the stereochemis-
try of 2. On the other hand, for the bound diamines, there is
a more pronounced difference between the RMS for the
correct isomer (i.e., the sorbitol congener) and its diaster-
eomers (DRMS=0.13–0.32), thereby resulting in a more
convincing identification.
Figure 3. Ensemble of the four main conformations of free diamine 2.
When diamine 3 is bound to 1, there are three best match-
ing conformers (1:1:2 molar ratio, RMS=0.48), one of
which (with ratio 2) is completely all-trans (Figure 4).
À
Within these, N N distances are 5.6, 6.6, and 7.6 ꢂ (all-
trans), thus indicating some flexibility of the clip 1, which
A more substantial gain by binding is obtained when
matching diamine 3 derived from xylitol against its diaster-
eomers. Here, the data for the free diamine were too degen-
erate to allow any analysis. For the bound diamine, which
was matched against experimental data for 3 bound to clip
1, a significant difference of RMS values was obtained, with
3 having the lowest value (3 (2R,3R,4S)=0.39, lyxitol conge-
ner (2S,3R,4S)=0.69, ribitol congener (2R,3S,4S)=0.71). We
suggest that this is mostly an effect of conformational rigidi-
fication of the bound molecules.
To summarize, these results indicate a significant improve-
ment in the stereochemical analysis on the basis of confor-
mational deconvolution by restricting the number of rele-
vant conformations through binding to a molecular clip.
Figure 4. Dominant conformations of bound 3.
À
can adjust the wall-to-wall (Zn Zn) distance depending on
Conclusion
the bound diamine. Previously, the equilibrium distance be-
tween Zn atoms in clip 1 was found to vary from approxi-
mately 7.0 to 12.4 ꢂ depending on the size of the ligand.
We have shown how the structural assignment of conforma-
tionally flexible molecules with several stereogenic centers
by NMR spectroscopy can be improved by binding to a sym-
metric, semirigid bisporphyrin molecular clip with two bind-
ing sites. An additional advantage of this host, besides rigidi-
fication of the guest molecule, is the separation of host sig-
À
Considering a Zn N bond length of typically 2.03 to
2.23 ꢂ,[19] this would correspond to ligand N N distances be-
À
tween 3.0 and 8.4 ꢂ being able to fit into the clip.
1
Distinction of stereoisomers: To investigate if different dia-
stereomers of a flexible molecule such as 2 could be distin-
guished from each other, we subjected its diastereomers to a
similar conformational deconvolution analysis (Table 2).
Conformational searches were performed for 2 and all of its
diastereomers (identified in Table 2 by the corresponding
parent alditols), excluding enantiomers. For the fitting pro-
cedure, experimental NOE or ROE values measured for
free 2 were used. The lowest RMS error was obtained for
nals, which otherwise might overlap in the H NMR spec-
trum. Furthermore, we also have devised a modified syn-
thetic
strategy
for
the
preparation
of
1,w-
diaminodideoxyalditol methyl ethers, starting from inexpen-
sive alditols. Initially formed dihydrochlorides are quantita-
tively transformed into the free bases by use of ion-ex-
change resin. The formation of cyclic ethers by demethyla-
tive cyclization can be avoided by minimizing the contact
with typical chromatographic stationary phases.
14636
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 14631 – 14638

Conformationally Flexible Diamines
FULL PAPER
H-4), 3.37 (ddd, 3J
3J
(H,H)=5.6, 5.4, 3.6 Hz, 1H; H-2), 2.99 (dd, J
H-1a), 2.96 (dd, 3J(H,H)=13.2, 4.0 Hz, 1H; H-6a), 2.85 (dd, 3J
13.5, 5.4 Hz, 1H; H-1b), 2.78 ppm (dd, 3J
(H,H)=13.2, 6.6 Hz, 1H; H-
ACHTUNGTRENUN(NG H,H)=6.6, 6.6, 4.0 Hz, 1H; H-5), 3.33 (ddd,
Experimental Section
3
G
ACHTUNGTRENNUNG
R
ACHTUNGTRENNUNG
Starting materials were purchased from commercial suppliers and were
used without purification. Room temperature (RT) refers to 20–228C.
Column chromatography was performed using Merck silica gel 60 (40–
63 mm), Fisher Florisil (60–100 mm), or Merck basic aluminum oxide (60–
200 mm). Thin-layer chromatography (TLC) was performed by using alu-
minum sheets precoated with silica gel 60 F254 (0.2 mm, E. Merck). Chro-
matographic spots were visualized by UV and, if required, by one of the
following stains: a solution of ninhydrin (2%) in ethanol, a solution of
H2SO4 (10%) in ethanol, or a basified solution of KMnO4 in H2O. MS-
ESI data were obtained using a Waters Micromass ZQ systems, EI,
70 eV; compounds were dissolved in methanol prior to analysis. High-res-
olution mass spectra were recorded using a Waters GCT Premier, CI
(methane), 70 eV, with a direct insertion probe. NMR spectra were re-
corded using Varian Unity Inova 500 MHz (1H 499.9 MHz, 13C
125.7 MHz), Varian Unity 400 MHz (1H 399.5 MHz, 13C 100.6 MHz), or
Varian Mercury Plus 300 MHz (1H 300.0 MHz, 13C 75.5 MHz) spectrome-
ters. Chemical shifts are referenced indirectly to tetramethylsilane
through the residual solvent signals (1H: CHCl3 at d=7.26 ppm, HDO at
d=4.79 ppm, 13C: CDCl3 at d=77.0 ppm). Signal assignments were de-
rived from 1H, 13C, COSY,[20] PE COSY,[21] gHSQC,[22] gHMBC,[23]
ROESY,[24] and TOCSY[25] spectra. For NMR spectroscopic titrations, ali-
quots of ligand solution were added to a solution of clip 1 in an NMR
spectroscopy tube. The relative stereochemistry of the synthesized com-
pounds was determined using 2D NMR spectroscopy and J-based config-
urational analysis.[1] Melting points were determined in open capillaries
using a Stuart Scientific SMP10 melting-point apparatus and are uncor-
rected. Specific optical rotations were measured using a Perkin–Elmer
241 polarimeter. Elemental analyses were performed by Eurofins Mikro-
kemi AB, Uppsala, Sweden.
AHCTUNGTRENNUNG
6b); 13C NMR (125 MHz, CDCl3, 258C): d=83.6 (C-2), 82.5 (C-5), 81.4
(C-4), 79.6 (C-3), 60.8 (OCH3-4), 60.5 (OCH3-3), 59.3 (OCH3-2), 57.6
(OCH3-5), 42.4 (C-6), 40.8 ppm (C-1); HRMS (CI): m/z calcd for
C10H24N2O4 [M+]: 237.1818; found: 237.1814.
Diamine 2 bound to clip 1: Clip 1 (3.1 mg, 0.0016 mmol) was dissolved in
Alox-filtered CDCl3 (0.6 mL) in an NMR spectroscopy tube. Aliquots of
2 dissolved in CDCl3 were added in small portions until a clip 1/2 molar
1
ratio of 1:0.8 was reached. H NMR (CDCl3, 500 MHz, 258C): d=1.67 (s,
3H; Me-4), 1.48 (s, 3H; Me-3), 1.43 (s, 3H; Me-2), 1.17 (s, 3H; Me-5),
0.69 (m, 1H; H-3), 0.57 (m, 1H; H-4), 0.04 (m, 1H; H-2), À0.22 (m, 1H;
H-5), À2.43 (m, 1H; H-6b), À2.58 (m, 1H; H-1b), À2.67 (m, 1H; H-6a),
À2.87 (m, 1H; H-1a), À4.27 (brs, 2H; NH2-C6), À4.35 ppm (brs, 2H;
NH2-C1); UV/Vis (CH2Cl2): lmax =428, 582, 672 nm.
AHCTUNGTERG(NNUN 2R,3R,4S)-1,5-Diamino-1,5-dideoxy-2,3,4-tri-O-methyl xylitol (3): Am-
berlite IRA-400 resin (chloride form, Sigma–Aldrich, 0.5 cm3) was rinsed
with H2O (5 mL), NaOH (1m, 8 mL) and H2O (5 mL) again, upon which
an AgNO3 test for ClÀ was performed to verify complete conversion of
the resin to the OHÀ form. This resin was rinsed with MeOH (3 mL) and
transferred into a 1 mL vial. (2R,3R,4S)-1,5-Diamino-1,5-dideoxy-2,3,4-
tri-O-methyl xylitol dihydrochloride (3h; 12 mg, 0.045 mmol) was dis-
solved in MeOH (0.5 mL) and poured over the resin. The mixture was
left to gently stir for 30 min under argon atmosphere and then filtered
through a glass wool plug. Evaporation of the solvent afforded 3 as pale
yellow oil (8.5 mg, 0.044 mmol, 99%). 1H NMR (500 MHz, CDCl3,
258C): d=3.53 (s, 3H; OCH3-3), 3.46 (s, 6H; OCH3-2+OCH3-4), 3.40
3
3
(dd, J
2H; H-2+H-4), 2.94 (dd, 3J
2.80 ppm (dd, 3J(H,H)=6.2, 13.3 Hz, 2H; H1-a+H5-a); 13C NMR
G
ACHTUNGERTN(NUNG H,H)=6.2, 4.8, 4.8 Hz,
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
Conformational analysis: Coupling constants were extracted from first-
order multiplets in 1H NMR spectra. NOE buildup experiments were per-
formed using NOESY[26] experiments with six different mixing times (see
the Supporting Information). NOE peak integrals were measured as an
average of both symmetry-related cross-peaks after thorough baseline
and phase corrections and normalized against both diagonal peaks. For
bound ligands, ROESY[24] experiments were performed with the same
strategy. Unrestricted conformational searches for each compound were
performed in MacroModel 9.9 with the OPLS-2005 force field and
CHCl3 as solvent to generate conformations that represented the entire
conformational space. This set of conformations was reduced by a redun-
dant conformer elimination using ConfGen[27] and/or manual elimination
of similar conformations to result in a set of 5–15 structures. For compari-
son of diastereoisomers in all-trans mode, the structure of 2 was built and
minimized, then chirality was inversed to create the other isomers from
(125 MHz, CDCl3, 258C): d=83.0 (C-3), 81.9 (C-2+C-4), 60.8 (OCH3-3),
58.9 (OCH3-2+OCH3-4), 42.3 ppm (C-1+C-5); HRMS (CI): m/z calcd
for C8H20N2O3 [M+]: 193.1541; found: 193.1539.
Diamine 3 bound to clip 1: Clip 1 (3.2 mg, 0.0016 mmol) was dissolved in
Alox-filtered CDCl3 (0.6 mL) in an NMR spectroscopic tube. Aliquots of
3 dissolved in CDCl3 were added in small portions until a clip 1/3 molar
ratio of 1:0.8 was reached. 1H NMR (CDCl3, 500 MHz): d=1.1 (s, 6H;
Me-2+Me-4), 1.09 (s, 3H; Me-3), À0.12 (dd, 3J
ACHTUNGTREN(NUNG H,H)=4.8, 4.8 Hz, 1H;
H-3), À0.26 (dm, 3J
(H,H)=4.2 Hz, 2H; H-2+H-4), À2.69 (m, 2H; H-
N
1b+H-5b), À2.94 (m, 2H; H-1a+H-5a), À4.41 ppm (brs, 4H; NH2);
UV/Vis (CH2Cl2): lmax =428, 582, 672 nm.
À
the same base structure. For ligands bound to clip 1, an N N distance re-
straint of 4 to 9 ꢂ was applied in a series depending on the ligand and
then combined. For the NAMFIS[3] analysis, DISCON[18] software was
utilized, using the goodness-of-fit expressed as an RMS deviation as an
indicator for agreement between calculated population of conformations
and experimental data, with lower values meaning better fit.
Acknowledgements
The Swedish Research Council is gratefully acknowledged for financial
support. We would like to thank Mꢃtꢄ Erdꢄlyi for valuable discussions
with regard to the conformational analysis, Thomas Norberg for advice
on carbohydrate chemistry, Prasad Polavarapu for preparation of the bis-
porphyrin clip, Vincent Andersson for initial synthetic studies of alditol
derivatives, and Johan Verendel for HRMS.
For the synthesis of 2a–h, 3a–h, and 4a–e, see the Supporting Informa-
tion.
(2S,3R,4R,5R)-1,6-Diamino-1,6-dideoxy-2,3,4,5-tetra-O-methyl sorbitol
(2): Amberlite IRA-400 resin (chloride form, Sigma–Aldrich, 0.5 cm3)
was rinsed with H2O (5 mL), NaOH (1m, 8 mL), and H2O (5 mL) again,
upon which an AgNO3 test for ClÀ was performed to verify complete
conversion of the resin to the OHÀ form. This resin was rinsed with
MeOH (3 mL) and transferred into a 1 mL vial. (2S,3R,4R,5R)-1,6-Di-
[1] G. Bifulco, P. Dambruoso, L. Gomez-Paloma, R. Riccio, Chem. Rev.
2007, 107, 3744–3779.
[2] a) C.-H. Tan, Y. Kobayashi, Y. Kishi, Angew. Chem. 2000, 112,
4452–4454; Angew. Chem. Int. Ed. 2000, 39, 4279–4281; b) Y. Ko-
bayashi, C.-H. Tan, Y. Kishi, J. Am. Chem. Soc. 2001, 123, 2076–
2078; c) Y. Kishi, Tetrahedron 2002, 58, 6239–6258; d) S. Higashi-
bayashi, W. Czechtizk, J. Am. Chem. Soc. 2003, 125, 14379–14393;
e) H. Seike, I. Ghosh, Y. Kishi, Org. Lett. 2006, 8, 3861–3864.
[3] a) D. O. Cicero, G. Barbato, R. Bazzo, J. Am. Chem. Soc. 1995, 117,
1027–1033; b) N. Nevins, D. Cicero, J. P. Snyder, J. Org. Chem.
1999, 64, 3979–3986; c) P. Dambruoso, C. Bassarello, G. Bifulco, G.
ACHTUNGTRENNUNGamino-1,6-dideoxy-2,3,4,5-tetra-O-methyl sorbitol dihydrochloride (2h;
9 mg, 0.03 mmol) was dissolved in MeOH (0.5 mL) and poured over the
resin. The mixture was left to stand for 30 min under an argon atmos-
phere and then filtered through a glass wool plug. Evaporation of the sol-
vent afforded 2 as a clear oil (7.1 mg, 0.03 mmol, 100%). [a]2D0 =31.9 (c=
8 in CH2Cl2). 1H NMR (500 MHz, CDCl3, 258C): d=3.53 (s, 3H; OCH3-
4), 3.50 (s, 3H; OCH3-2), 3.48 (s, 3H; OCH3-3), 3.47 (dd, 3J
3.4 Hz, 1H; H-3), 3.42 (s, 3H; OCH3-5), 3.42 (dm, 3J
(H,H)=3.4 Hz, 1H;
ACHTUNGTRENUN(NG H,H)=5.6,
AHCTUNGTRENNUNG
Chem. Eur. J. 2013, 19, 14631 – 14638
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
14637

A. Gogoll et al.
Appendino, A. Battaglia, G. Fontana, L. Gomez-Paloma, Org. Lett.
2005, 7, 983–986.
[13] a) V. GlaÅon, A. El Meslouti, R. Uzan, G. Demailly, D. Baupꢇre,
Tetrahedron Lett. 1996, 37, 3683–3686; b) M. Mancera, I. Roffꢄ, M.
Rivas, C. Silva, J. A. Galbis, Carbohydr. Res. 2002, 337, 607–611;
c) V. GlaÅon, M. Benazza, A. El Anzi, D. Beaupꢇre, G. Demailly,
Carbohydr. Chem. 2004, 23, 95–110.
[14] a) D. E. Ames, R. E. Bowman, J. Chem. Soc. 1952, 1057–1068; b) L.
Toom, A. Kꢈtt, I. Kaljurand, I. Leito, H. Ottosson, H. Grennberg,
A. Gogoll, J. Org. Chem. 2006, 71, 7155–7164; c) K. Muranaka, A.
Sano, S. Ichikawa, A. Matsuda, Bioorg. Med. Chem. 2008, 16, 5862–
5870.
[4] a) P. Thepchatri, D. O. Cicero, E. Monteagudo, A. K. Ghosh, B. Cor-
nett, D. C. Liotta, J. P. Snyder, J. Am. Chem. Soc. 2005, 127, 12838–
12846; b) J. J. Koivisto, E. T. T. Kumpulainen, A. M. P. Koskinen,
Org. Biomol. Chem. 2010, 8, 2103–2116; c) H. Andersson, H. De-
maegdt, G. Vauquelin, G. Lindeberg, A. Karlꢄn, M. Hallberg, M. Er-
dꢄlyi, A. Hallberg, J. Med. Chem. 2010, 53, 8059–8071; d) M.
Fridꢄn-Saxin, T. Seifert, L. K. Hansen, M. Grøtli, M. Erdꢄlyi, K.
Luthman, Tetrahedron 2012, 68, 7035–7040.
[5] a) R. J. Abraham, G. R. Bedford, B. Wright, Org. Magn. Reson.
1982, 18, 45–52; b) R. J. Abraham, C. J. Medforth, Magn. Reson.
Chem. 1987, 25, 432–438; c) R. J. Abraham, C. J. Medforth, Magn.
Reson. Chem. 1990, 28, 343–347; d) K. J. Cross, M. J. Crossley, Aust.
J. Chem. 1992, 45, 991–1004.
[15] F. J. Feher, K. D. Wyndham, D. Soulivong, F. Nguyen, J. Chem. Soc.
Dalton Trans. 1999, 1491–1498.
[16] a) G. R. Gray, F. C. Hartman, R. Barker, J. Org. Chem. 1965, 30,
2020–2024; b) H. Dehmlow, J. Mulzer, C. Seilz, A. R. Strecker, A.
Kohlmann, Tetrahedron Lett. 1992, 33, 3607–3610; c) O. R. Martin,
F. Yang, F. Xie, Tetrahedron Lett. 1995, 36, 47–50; d) B.-H. Yang, J.-
Q. Jiang, K. Ma, H.-M. Wu, Tetrahedron Lett. 1995, 36, 2831–2834;
e) R. Persky, A. Albeck, J. Org. Chem. 2000, 65, 3775–3780; f) R.
Cribiꢉ, I. Cumpstey, Chem. Commun. 2008, 1246–1248; g) M.
[6] a) J. N. H. Reek, A. P. H. J. Schenning, A. W. Bosman, E. W. Meijer,
M. J. Crossley, Chem. Commun. 1998, 11–12; b) M. J. Crossley, P.
Thordarson, Angew. Chem. 2002, 114, 1785–1788; Angew. Chem.
Int. Ed. 2002, 41, 1709–1712; c) N. Solladiꢄ, F. Aziat, S. Bouatra, R.
Rein, J. Porphyrins Phthalocyanines 2008, 12, 1250–1260.
[7] a) T. Hayashi, M. Nonoguchi, T. Aya, H. Ogoshi, Tetrahedron Lett.
1997, 38, 1603–1606; b) X. Huang, K. Nakanishi, N. Berova, Chirali-
ty 2000, 12, 237–255; c) X. Huang, N. Fujioka, G. Pescitelli, F. E.
Koehn, R. T. Williamson, K. Nakanishi, N. Berova, J. Am. Chem.
Soc. 2002, 124, 10320–10335; d) A. G. Petrovic, G. Vantomme, Y.
Negrꢅn-Abril, E. Lubian, G. Saielli, I. Menegazzo, R. Cordero, G.
Proni, K. Nakanishi, T. Carofiglio, N. Berova, Chirality 2011, 23,
808–819; e) M. Tanasova, B. Borhan, Eur. J. Org. Chem. 2012,
3261–3269.
ˇ
Cerny, Adv. Carbohydr. Chem. Biochem. 2003, 58, 121–198.
[17] J. Defaye, D. Horton, Carbohydr. Res. 1970, 14, 128–132.
[18] a) O. Atasoylu, G. Furst, C. Risatti, A. B. Smith, Org. Lett. 2010, 12,
1788; b) http://discon.sourceforge.net.
[19] C. B. Storm, J. Am. Chem. Soc. 1970, 92, 1423–1425.
[20] a) A. Wokaun, R. R. Ernst, Chem. Phys. Lett. 1977, 52, 407–412;
b) A. J. Shaka, R. Freeman, J. Magn. Reson. 1983, 51, 169–173.
[21] L. Mueller, J. Magn. Reson. 1987, 72, 191–196.
[22] A. L. Davis, J. Keeler, E. D. Laue, D. Moskau, J. Magn. Reson. 1992,
98, 207–216.
[8] S. Norrehed, P. Polavarapu, W. Yang, A. Gogoll, H. Grennberg, Tet-
rahedron 2013, 69, 7131—7138.
[9] a) J. A. Galbis, M. G. Garcꢆa-Martꢆn, Top. Curr. Chem. 2010, 295,
147–176; b) M. G. Garcꢆa-Martꢆn, E. B. Hernꢃndez, R. R. Pꢄrez,
J. A. Galbis, Polym. Degrad. Stab. 2008, 93, 1370–1375.
[10] a) M. G. Garcꢆa-Martꢆn, E. B. Hernandez, R. R. Perez, A. Alla, S.
Munoz-Guerra, J. A. Galbis, Macromolecules 2004, 37, 5550–5556;
b) M. Mancera, F. Zamora, I. Roffe, M. Bermudez, A. Alla, S.
Munoz-Guerra, J. A. Galbis, Macromolecules 2004, 37, 2779–2783.
[11] M. G. Garcꢆa-Martꢆn, R. R. Pꢄrez, E. B. Hernꢃndez, J. A. Galbis,
Carbohydr. Res. 2001, 333, 95–103.
[23] R. E. Hurd, B. K. John, J. Magn. Reson. 1991, 91, 648–653.
[24] A. Bax, D. G. Davis, J. Magn. Reson. 1985, 63, 207–213.
[25] L. Braunschweiler, R. R. Ernst, J. Magn. Reson. 1983, 53, 521–528.
[26] a) A. Kumar, R. R. Ernst, K. Wꢈthrich, Biochem. Biophys. Res.
Commun. 1980, 95, 1–6; b) G. Bodenhausen, H. Kogler, R. R.
Ernst, J. Magn. Reson. 1984, 58, 370–388.
[27] K. S. Watts, P. Dalal, R. B. Murphy, W. Sherman, R. A. Friesner,
J. C. Shelley, J. Chem. Inf. Model. 2010, 50, 534–546.
Received: February 10, 2013
Revised: June 6, 2013
Published online: September 11, 2013
[12] R. M. De Lederkremer, C. Marino, Adv. Carbohydr. Chem. Bio-
chem. 2003, 58, 199–306.
14638
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 14631 – 14638